Simulation model of a coal bulk terminal

Customer: JSC «Daltransugol», a subsidiary of JSC «SUEK».
Task: Study of the Terminal operation and calculation of its capacity and individual components within the framework of the project «Increasing the transshipment capacity of JSC «Daltransugol» to 40 million tons of coal per year.

Briefly about the object

The Vanino Bulk Terminal is the most important transshipment point on the way to the markets of the Asia-Pacific region, located in the deep-water Muchke Bay and located at the extreme point of the Baikal-Amur Mainline, is one of the youngest and most modern coal terminals in Russia. It has direct access to two independent railway lines - Trans-Siberian and Baikal-Amur, which, in turn, connect the port with the entire territory of Russia.

  1. The terminal is equipped with an automated wagon unloading system and a coal warehouse with a capacity of up to 1.2 million tons.
  2. The current pier is capable of receiving and handling Capesize vessels.
  3. At the end of 2020, the cargo turnover amounted to 23 million tons of coal.
  4. Automated conveyor network with a length of more than 4 km.
  5. The length of railway tracks is more than 50 km. 

Simulated processes

  1. Formation of applications for the supply of goods to the terminal by rail and applications for the shipment of coal by bulk carriers.
  2. Movement of trains at the adjacent station and internal railway logistics of loaded wagons: arrival, sorting, sawing and defrosting of cargo, filing for unloading.
  3. Unloading of wagons by means of an automatic system of wagon dumpers, taking into account the peculiarities of the process at different times of the year and weather conditions.
  4. Transportation of goods through the conveyor network.
  5. Simulation of the work of a virtual dispatcher for marking a warehouse for unloading coal.
  6. Building routes, taking into account the priorities of cargo, coal grades, season, weather, workload and degree of operating time of the main technological equipment, etc.
  7. Loading of coal onto ships, taking into account their deadweights and restrictions on individual berths for receiving ships; queue priorities, as well as types of vessels for which loading should proceed with different quality control and speed, etc.
  8. Movement of empty wagons along the internal railway network, including sorting, rejection of wagons and sending them for repair, formation of trains for shipment by rail. 

End-to-End Business Process Modeling Boundaries 

Entry point: arrival of laden wagons at the Terminal Reception Park.
Exit point: departure of empty wagons after the Terminal Sorting Park and departure of loaded vessels from the water area. 


The simulation model development process consisted of the following stages: 

  1. Preparing for modeling
    At this stage, the specialists of Dilibrium LLC conducted a study of the modeling object - the Vanino Bulk Terminal - with the departure of specialists to the object. The parameters of the technological process were specified, full-scale measurements of individual technological operations were made, interviews were conducted with various specialists and terminal services, historical data on the operation of the terminal were collected and analyzed, as a result of which statistical patterns were found, which were subsequently used to build a model.

  2. Model design
    At this stage, information about the object of modeling was summarized and the concept of a simulation model was formed, in which the goals and objectives of modeling were clarified; modeling boundaries are defined; the structure and technological architecture of the simulation model, as well as the main set of capabilities that the simulation model implements and the simulation results are described.

  3. Development of the model «AS IS»
    At the design stage of the simulation model, it was decided to first develop the simulation model of the terminal in the existing development scheme «AS IS». This was important in order to set sufficiently accurate values of the parameters of individual agents and algorithms on the constructed and sufficiently detailed model in order to verify the model in accordance with historical data. As a result of the verification of the terminal simulation model, it was possible to achieve high reliability in comparison with historical data for 2019-2020.
    This result became possible due to the high degree of detail of technological processes in the simulation model.
    To simulate technological processes and change the state or behavior of model objects, the following simulation methods were applied, namely:
    • a method of agent-based simulation to implement the behavior of individual agents;
    • discrete-event modeling method for modeling the technological processes of the Terminal.

  4. Refinement of the model with «TO BE» scenarios
    To the verified «AS IS» simulation model with fine-tuned processes and parameters, several experiments with modified terminal spatial planning, new layout and new types of main technological equipment were added to test hypotheses for calculating the maximum throughput of the terminal. In total, 5 different options for the configuration of warehouses and sets of basic technological equipment were implemented in the model, corresponding to various options for the future development of the Terminal. 

To solve the problem, a simulation model was developed in the «Anylogic» software environment. With the help of the «Anylogic Railway Library», all internal railway logistics was modeled with a fairly extensive system of tracks and technological processes. The pipeline network was developed on the basis of the stream library. Additional technological processes were modeled using Java code.

In addition to modeling railway and conveyor logistics, it was quite a difficult task to develop a cargo routing algorithm taking into account warehouse planning, since a large number of interrelated conditions and parameters had to be taken into account in order to select and build a route:

  1. the stage of modernization chosen before the simulation;
  2. weather conditions (were modeled both according to historical data and with specified coefficients of deviations);
  3. season;
  4. availability of each key unit = equipment (more than 80 units): employment on the line, maintenance and repair, breakdown;
  5. current order from the dispatcher for unloading or loading;
  6. volume and types of cargo in wagons at the Terminal and on the way;
  7. volume and brand of cargo in the Warehouse;
  8. current marking of the coal stored in the warehouse, in certain Stacks;
  9. operation priority (unloading/loading);
  10. deadweight of bulk carrier in queue for loading;
  11. etc.

This algorithm was developed using the «Anylogic library - State Diagram».

The model simulation period took into account a full calendar year (365 days), including seasonality and weather conditions. During this period, both operational data were collected in virtual time mode, as well as statistics by year/month/day for all major processes and pieces of equipment. At the end of the simulation, the final statistics output was uploaded to a separate excel file for further analysis of the results of the experiments.
Stochastics were taken into account in the model, for example, random deviations from statistical data on breakdowns / weather conditions, or the coefficient of errors / inconsistencies when submitting trains and ships to the Terminal due to the human factor, unscheduled changes in the Russian Railways schedule, changes on the part of the Customer, cargo, weather. Conventionally, a ship could come to the pier, for which not all of the required grades of coal were in stock, and the rest was still delayed on the way to the Terminal. 


The simulation model was verified through the «AS-IS» stage, and with the help of multiple experiments for each variant of the modernization stage, extensive statistics were obtained both for maximum productivity and for the indicators of each equipment node (maintenance and repair, operating hours, CTG, KPI and other values), which in the model are more 80 units. Also, with the help of the experiments, bottlenecks for different input parameters were identified.
St. Petersburg.